dots.ocr vs OpenOCR
Pick dots.ocr for multilingual PDFs where you need layout boxes, tables as HTML and formulas as LaTeX from one 3B model. Pick OpenOCR for lightweight self-hosted parsing of Chinese and English documents, since its 0.1B models report 90.57 on OmniDocBench v1.5. The main split is language coverage and size against output richness.
dots.ocrrednote/dots-ocr | OpenOCRfudan/openocr | |
|---|---|---|
| Type | Open source | Open source |
| Price | 2 credits / page | 1 credit / page |
| Per 1,000 pages | $2.00 | $1.00 |
| Typical speed | 1.5 s / page | 0.5 s / page |
| Best for | Multilingual document parsing | Lightweight OCR pipeline on CPU |
| Inputs | PDF, images | PDF, images |
| Languages | 100+ languages | Chinese and English |
| Max pages per request | 500 | 500 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | No | No |
| Math | ✓ Yes | ✓ Yes |
| Bounding boxes | ✓ Yes | ✓ Yes |
| Confidence | No | No |
| Forms | No | No |
| Multilingual | ✓ Yes | No |
| On OCRRouter | Available | Available |
Prices are what you pay on OCRRouter. Speed is the provider's typical figure, not a benchmark we ran.
Which to choose
Choose dots.ocr if
- You need more than Chinese and English, since dots.ocr is described as multilingual while OpenOCR supports only those two.
- You want JSON, Markdown, HTML, LaTeX and bounding boxes from a single model, with tables as HTML and formulas as LaTeX.
- You want chart and diagram conversion to SVG through the dots.mocr-svg variant.
Choose OpenOCR if
- You work with Chinese and English academic PDFs with formulas and want a 0.1B model instead of a 3B one.
- You want the Apache-2.0 licence, with ONNX export for deployment.
- You do scene text recognition research or need detection and recognition in one toolkit.
Key differences
| dots.ocr | OpenOCR | |
|---|---|---|
| Licence | MIT (the repository also includes a separate dots.ocr license agreement file) | Apache-2.0 |
| Model size | 3B (1.7B LLM backbone) | 0.1B (OpenDoc-0.1B and UniRec-0.1B) |
| How you run it | Self-hosted via vLLM or Hugging Face Transformers; a single CUDA GPU with vLLM is the documented path | Self-hosted as a Python library; no hardware requirements or throughput figures published |
| Output formats | JSON, Markdown, HTML, LaTeX, bounding boxes, SVG | Plain text, LaTeX, bounding boxes |
| Languages | Multilingual | Chinese and English only |
| Benchmark (author-reported) | olmOCR-bench 79.1 (dots.ocr), 83.9 (dots.mocr 1.5) | OmniDocBench v1.5 overall 90.57 (OpenDoc-0.1B) |
| Known failure | Runs of ellipses or underscores can make output repeat endlessly; tables above 1000 grid cells raised an error and dropped the page in a Docling report | Document parsing depends on a separate layout model, PP-DocLayoutV2 from PaddlePaddle |
Sources
What each one is
dots.ocr
Single vision model from rednote that does layout detection and content recognition together.
dots.ocr detailsOpenOCR
Fudan's open toolkit for text detection and recognition, formulas, tables and document parsing.
OpenOCR detailsTry both on your own file
Run the file once with each model and compare the output.
PDF, PNG, JPG, WebP, HEIC. Up to 20 MB each, 10 files.No file?
Use one, fall back to the other
{"model": "rednote/dots-ocr","fallbacks": ["fudan/openocr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
dots.ocr vs RolmOCRdots.ocr vs Tesseract.jsNanonets-OCR vs OpenOCRSurya vs OpenOCRQianfan OCR vs OpenOCRHunyuan OCR vs OpenOCRDonut vs OpenOCRPaddleOCR-VL vs OpenOCR
dots.ocr alternativesOpenOCR alternativesBest OCR models in 2026